JBoss Founder Interview

peterdaly writes "The JBoss website has an interview with Marc Fleury, the JBoss founder regarding his vision. In case you have been living under a rock, JBoss is an Open Source Java Application Server (J2EE) which has been picking up tons of steam recently, especially with the recent introduction of features like clustering. Competing products from companies like IBM (WebSphere) and BEA (WebLogic) go for tens of thousands of dollars, which is interesting since JBoss is starting to have features the big boys don't. JBoss had 72,000 downloads in October. This is a project to watch."

Comments

C#?

[Anonymous Coward]

How far do you think Java can go from here? What are the markets that it can likely succeed in the future?

What exactly is a Java application server?

[svara]

I'm not really following JAVA as I'm more interested in "traditional" programming languages - but this sounds interesting? What's a Java application server? And it has clustering? So it's some kind of physical server computer running Java apps?

Re:What exactly is a Java application server?

[spackled42]

A Java application server is a platform written primarily in Java that serves web content through Servlets. For example, IBM has WebSphere Application Server (WAS) that is the core to most of their web products. It is a servlet manager with a DB API all combined with a modified version of Apache. You can develop with this directly, or you can purchase the next step up of IBM Websphere Commerce Suite which has tools and a DB schema for web (especially e-commerce) development.

An app server is not a physical computer, merely a server package akin to a web server or ftp server. The clustering capabilities refer to the capability to share the load across several machines and having been programmed with parallel computing in mind.

--Dave

Re:What exactly is a Java application server?

[iapetus]

Not strictly right, AIUI.

The Websphere application family is built up on several layers (described by IBM as the e-business application framework). The Foundation layer contains those tools that provide the fundamental low-level functionality of the system, and this basically comprises WAS and MQ Series.

On top of this you've got the Foundation Extension layer: tools for optimising the performance of the Foundation tools, and for development and deployment. Here you've got tools like VisualAge for Java, the personalization suite for Websphere, and Edge Server (used to be called the performance pack, IIRC, but clearly that didn't sound cool enough...)

Finally, over the top of this you have the Application Accelerator layer, consisting of tools for building particular types of application. That's where Commerce Suite lives, along with tools like the Websphere B2B Integrator.

JBoss itself corresponds to just part of the Websphere Application Server - it's really just an EJB server rather than a full J2EE server. However, combined with Tomcat (and you can download the two as a nicely integrated bundle from the JBoss site) you end up with something comparable in many ways with Websphere Application Server (advanced edition). In some ways it's better, in others it's worse. It's certainly more up-to-date - Websphere is (as with most IBM Java systems) a couple of versions behind the bleeding edge. That gives it the benefit of stability and reliability, but it gives JBoss the advantage in enhanced functionality (such as support for CMP2.0).

Re:What exactly is a Java application server?

[igbrown]

A more accurate description is a platform that handles "back-end" functions for distributed or networked applications. This might include accessing databases or performing calculations or functions in a centralized location. It is not limited to web-based applications, though this is a primary use for application servers. The generic model is for various clients (browsers, standalone applications, or some type of service) calling functions that are executed by the application server. Often, information is then passed back to the client. The domain associated with application servers is the execution of "business logic". Servlets are often part of an application server platforms but application servers are not limited to serving up dynamic web content.

Re:What exactly is a Java application server?

[felipeal]

IBM has WebSphere Application Server (WAS) that is the core to most of their web products. It is a servlet manager with a DB API all combined with a modified version of Apache

Just a small correction: even though WAS provides its own version of apache, it also provides a plug-in to be used with the original apache

Most of the servlet engines out there (like Tomcat, JRun and WebSphere) offer the option to integrate with an external web server, mainly because they don't perform as well as the "regular" web servers on static pages.

Re:What exactly is a Java application server?

[Anonymous Coward]

God, you are ignorant.

Java1.1 maybe, but Java2? If slashdot was built using J2EE, it'd never slow down and need to go to straight HTML from time to time.

J2EE has proven, time and time again, to stomp on any Perl/CGI webpage (even with mod_perl).

Re:What exactly is a Java application server?

[iapetus]

I believe it's actually the same plugin for the IBM HTTP Server and Apache - the differences between them don't affect the interaction with WAS. There are also plug-ins for a range of other servers, including IIS, Netscape Enterprise Server, Domino, and Domino Go Webserver.

Re:What exactly is a Java application server?

[Fnkmaster]

JBoss is definitely not just an EJB server. It provides a JMX managed component framework with components that wrap around the various services that comprise a J2EE app. It provides JNDI, JMS, EJB 2.0, JMX, etc. I uses third party implementations of the Servlet/JSP subsystem which is appropriate, there's really no need to have anything more tightly integrated than that. JBoss also has a JCA implementation, obviously a JTA/JTS implementation, JAAS implementation, and is SOAP capable. So I think it is pretty complete. :)

Re:What exactly is a Java application server?

[iapetus]

True - I was oversimplifying horribly there. :) Just trying to separate the functionality in JBoss proper (EJB, but also, as you point out, JNDI, JMS, and other related services) from the Servlet/JSP functionality provided by external apps (Jetty/Tomcat) which are required for a full J2EE solution.

Re:What exactly is a Java application server?

[Anonymous Coward]

Drank the kool-aid, did you?

Anytime I see those 3 magic letters "jsp" I know to expect painfully slow response times. And on the back end, 2-3x more servers per transaction rate.

Re:What exactly is a Java application server?

[Anonymous Coward]

Ok, this is probably dumb, but I gotta ask it:

What is the difference between this and a PHP/Apache server? If I replaced my single PHP/Apache/MySQL server with a Java application server, what would that give me?

Re:What exactly is a Java application server?

[Anonymous Coward]

what the FUCK are you trying to say?
I'm going to enter that reply in the international obfuscated english contest.

Ohh my Hemos..

[Count]

"..picking up steam"

..Yeah It is also a real boost for the Open Source movemnet. This project has been brewing and is sure to mug the competition.

heheh

Re:Ohh my Hemos..

[British]

Don't look at me. Apparently I have been living under a rock for not knowing this java dude's name.

Tomcat

[RazzleFrog]

How does this compare with Apache's Tomcat (Jakarta) project?

Re:Tomcat

[Zach Garner]

Tomcat is only a basic application server (if it can be considered one at all) and does not support J2EE.

Re:Tomcat

[jeffy124]

Tomcat is for Servlets, not a J2EE app server.

Re:Tomcat

[felipeal]

He doesn't compete with Tomcat, it actually even uses it.

Tomcat is not a J2EE application server, it's basically a jsp/servlet engine.

An appserver has many more features, like JNDI, transaction, connection pooling, EJB, etc...

Re:Tomcat

[RazzleFrog]

Thanks for the great posts. I've just started working with JSP/Servlets so Tomcat is what I've been using.

Re:Tomcat

[snoopdave]

Actually, Tomcat is a J2EE compliant "Web Container" so it must support Servlets and JSP, but it does not have to support EJB, JNDI, JMS, etc. etc.

Re:Tomcat

[GrayArea]

Actually, Tomcat 4.0 has its own JNDI provider, so it does support looking up resources (JDBC datasources, etc.) in a standard way. It is also possible to extend its JNDI provider to support your custom resources, they have an example where you can register an arbitrary Java Bean in your web app, and then look it up from a servlet.

Re:Tomcat

[MSBob]

Cool. So once you've built your first web application and discovered the headaches of implementing your own persistence, transactions, messaging etc. you'll be able to understand the value JBoss brings to the table :).

But how does it run

[tcd004]

Perhaps if MS had not shurgged off serious Java development so early in thier OS cycle, they could join in on this buttery goodness. Now you can only buy software like this for windows from companies more interested in labeling themselves "enterprise" than producing affordable, useful stuff.

But how does it run on a Pentium?

Re:But how does it run

[Anonymous Coward]

To my knowledge, Java tends to run sweet like candy for web servers, since it doesn't have to start a new process for every single page update. Maybe someone else will clarify, but just in case they don't, I'll just say that in the general case, Java isn't the grindingly slow language that some people seem to think it is..

To Watch?

[md17]

This is not just a project to watch... When 3.0 is release, all other Java Middleware will be worthless. It is more than a project to watch, but one to support, use, and contribute to.

Re:To Watch?

[Anonymous Coward]

Yeah, cause the other products aren't being expanded also.

Honestly, if you are getting competition from an open source project, tell me you won't be tracking their every move (and know exactly what they are trying to release).

How many developers for JBoss are undercover for IBM and BEA?

Think about it...

Re:To Watch?

[FortKnox]

all other Java Middleware will be worthless

Now there is a claim!
What about the partnerships and certifications from BEA and IBM? All worthless?

Remember there are two facets of software. The reliability/overall efficiency of the software, and the BUSINESS of the software.

Linux is a GREAT operating system, but MS has the better grasp of business.
I'm NOT trying to flame or start an OS war, just showing that business plays as important a role as the overall product itself...

Re:To Watch?

[iapetus]

Bingo. IBM's offerings aren't just about the app server - that's actually a very small part of the Websphere family of products. It's the development tools, the extra features, the integration with other IBM back-end systems that make Websphere so attractive, not just the ability to run EJBs. If that's what appeals to you for a particular project, then JBoss can't come close to matching it right now.

That said, if you just want the basic J2EE functionality, you could do a lot worse than JBoss, for a lot more money.

Re:To Watch?

[Fnkmaster]

I'm not going to disagree on this entirely - if you are willing and able in a shop to go the pure IBM route, then you can get a lot of benefits from that approach. However, you suffer from SERIOUS vendor tie in. Compared more realistically to a product like Weblogic, however, I think JBoss stands up admirably. You can get integration with IDEs and debuggers via add-on modules to JBuilder, etc. It may not be as easy to get and as smooth as the pure IBM approach, but you can maintain some degree of portability and you can keep your costs drastically lower, and still deploy in non-IBM shops. Anyway, just my experiences as the CTO of a company making J2EE-based apps that we sell to others - we can't afford to be an IBM shop in every sense of that word.

Re:To Watch?

[iapetus]

And you wouldn't want to be an IBM shop in that sort of situation: firstly you've got to be able to sell to people who don't have a 100% IBM infrastructure, and secondly it's not always the right tool for the job. I'm just saying that there are extra features available with the commercial tools and that JBoss isn't currently a complete replacement for them. It's concentrating on its core functionality, and as far as I'm concerned (from a selfish point of view) that's great - it fits perfectly with what I want to use it for.

Re:To Watch?

[Harmast]

The real benefits of the pure IBM solution are going to be in one (well two) main area: financial and insurance companies using IBM big iron.

That's what my company has. While normally I would have fought for JBoss/Tomcat for our Ebusiness the Websphere application setup just adds too many useful things to integrate into our legacy systems. One thing is the ability to roll IMS transactions into a bundle with DB2 and Oracle transactions as a single supertransaction. To understand how powerful this is, if IMS transaction two requires transaction one to commit before running and Oracle transaction A requires IMS one/two to commit running you can still run all three and then rollback ALL of them when the last one fails (if you're wondering why they weren't written that way, remember these were written for step by step 3270 usage, not one webform). That alone makes paying for Websphere worthwhile for IBM big iron shops.

Those who use a server network and RDMS for the majority of their processing, though, are better off with JBoss (usually).

Cool stuff

[Putz19]

I think this is a great project. We use Websphere where I work, and that product is cool to work with. The bigest problem is the price. I just hope they can get the program to be as good/BETTER then the expensive ones.

"Things like Dynamic proxies with compilation-less deployment have made us the darling of serious development shops"

Already showing it does stuff Websphere can not. How SWEET!!

The Apache of Java servers?

[imrdkl]

This sounds pretty nice. We've already sold our souls to BEA, but having used Resin and Tomcat for production deployment before BEA, we are not afraid to try something new. BEA is still not doing for us what we expected, especially in it's loadbalancing and proxying capabilities. The Apache integration for BEA requires a new vhost for every single cluster, and the "load balancing" is simple round-robin. It sounds like this JBoss has that beat already.

An apache integration is important for us, however, since apache is our frontline proxy. Anyone know if JBoss has an apache module yet?

Re:The Apache of Java servers?

[aim4min]

Quick answer to your question -- JBoss runs within an APP server. So you run JBoss within Resin or Tomcat. You don't need anything to connect JBoss with your Webserver.

Re:The Apache of Java servers?

[peterdaly]

JBoss has a version which include a version of Tomcat, which runs inside the same JVM as the app server, which improves the performance greatly in many cases. That means no network connection, even locally, for RMI calls. Tomcat itself (Jakarta really), Apache's JSP/Servlet container, even as a seperate product integrates very well into Apache since they are made to work together.

Don't know if that answers your question or not.

-Pete

Re:The Apache of Java servers?

[Brummund]

Yes, JBoss integrates with both Jetty and Tomcat.
Jetty is full-fledged http server, and Tomcat has an Apache module.

You can download JBoss/[Tomcat|Jetty] bundles, and lo and behold, they run "out of the box" on both Linux and Windows.

JBoss is extremely convenient for development since it doesn't require any special compilers, converters or whatever BEA call their ejbc-tools; just jar it up, drop it in the 'jboss/deploy' directory and watch the log file.

With the new versions of Ant supporting ear and war in addition to ordinary jars, the build.xml (ie. the Makefile) for for a JBoss project is dead easy to write.

Several IDEs are now supporting JBoss as well, providing support for those "corporate developers" stuck in "tool hell" :)

(I use Emacs and Ant and would only swap those for a very fat paycheck :)

Re:The Apache of Java servers?

[atomray]

I don't think that's the answer. Many of us would still like to use Apache in front of Tomcat. This is easy to do with Apache + Tomcat when downloaded from Apache.org, but the integrated JBoss/Tomcat downloads from JBoss don't integrate easily with Apache (tomcat's config files are ignored when instatiated by JBoss).

I've looked into this but haven't had the time to make it work. Go to JBoss.org and go through the forums, people have made this work, and there are vague instructions :).

Performance?

[betis70]

I have heard a number of negative things about JBoss's performance under heavy loads. Anyone with experience using it care to comment? It seems like the clustering would really help such situations, and I am excited to see it advancing as well as it is. ..

Re:Performance?

[Fnkmaster]

I believe what you are referring to is Servlet performance. If you are using an embedded servlet engine in production for a high volume site and servlet processing seems to be partially responsible for the performance problems, then I would recommend using embedded Jetty rather than Tomcat (I see in Rickard's post that he can't, but that's his problem, and not a usual one).

Honestly, most of the J2EE apps I've seen, Servlet performance is a MINISCULE problem compared to much, much larger application architecture issues. You can always add more servlet engines on separate boxes, of course, then the RMI invocations are going to be slow. You did pay attention to using _coarse_ grained components, didn't you? Anyway, this is just my opinion.

In any case, you can't blame the JBoss team for this - you can pretty much embed any servlet engine you want into JBoss - Tomcat and Jetty are pre-done for you, and if some enterprising hackers want to find a better solution and embed that, it isn't that hard to do.

Re:Performance?

[icoloma]

The various comparisons I have seen rate it down only because you can put weblogic in a cluster but not JBoss.

Well, now you can. JBoss 3.0 fully supports two kinds of clustering: farms of servers and true clusters.

Re:Performance?

[waveclaw]

During a two semester introductory course in software engineering I participated in a group that refractored a J2EE application (improved underlying code yet left it bug-for-bug compatible) and then completely reimplementation another J2EE application concept. Throughout both experiences, the rich support of architectures and patterns, on which JBoss (and Tomcat) are designed, made the projects simpler and more effective at achieving the goals of writing maintainable code.

More importantly for this point-of-view is these (like many other programming techniques) allowed the group to implement solutions with widely varying performance: from slow but explicit code to hand-tooled solutions that utilize aspects of the problem domain. While it is a common conception that systems with as rich a feature set as JBoss are performance hogs, it should be noted that most enterprise scale systems don't have very attractive small-scale performance (some Oracle DBs can have a +1GB footprint before any user data or DBs are added which ain't exactly embedded.) JBoss is both small and fast enough to run from student/school-lab quality development systems (real budget machines IHMO) but has sufficient performance to match wits with a full-blown ecommerce application. I helped implement a complete on-line store (liquor store, that is) that held under our test loads (page view/transactions per sec) without relying on clustering.

The onjava.com article does bring up a valid point: like many complex software packages (tcsh, ant, Java) does have a complex environment to setup and maintain. A large part of this however (much like with tcshrc config files) it is a matter of how much you want to custom tailor YOUR system. Much like makefiles and easy builds, a little bit o' work goes a long way (I've seen a factor of ten improvement with some cases.) I think though you will find the best improvement in performance not from hoping between J2EE systems but from good ol' fashion Data Structures and Algorithm Analysis appropriate to you application.

Happy coding.

Re:Performance?

[Anonymous Coward]

Have you seen any EJB application with good performance?

Jakarta Plug & My AppServer Experiences

[FortKnox]

...JBoss article.... must... plug... Jakarta Project.
Jakarta contains whole bunches of open source tools that work great for Java Projects (I'm using struts and ant on my current project).
They all work extremely well (and simple to install) with JBoss.

I don't know the level of people using JBoss, though. The top two app servers are WebSphere and Weblogic. They take 50% of the market. The next is iPlanet (netscape), then I think its JBoss. So, even though its the cheapest (free), doesn't mean its got the market.
It'll be tough to crack WebSphere & WebLogic.
What JBoss needs is a certification (with levels) for developers to obtain.
If I go to a client and say "I have a level 3 WebLogic certification, a level 2 WebSphere certification, and know JBoss", what are they gonna pick?

Re:Jakarta Plug & My AppServer Experiences

[Kerg]

Do notice though, that many of the "analysts" measure the market by direct license revenue

Independent research by Gartner Dataquest shows that BEA holds 41% of the new direct license revenue in the application server software market in 2000, a figure that is 10 share points higher than its closest competitor.

The Gartner market analysis revealed the following market share results: BEA - 41% IBM - 31% iPlanet - 13% Oracle - 4% HP Bluestone - 4%

JBoss does not show up here because the server code itself does not generate any revenues.

Re:Jakarta Plug & My AppServer Experiences

[Kingpin]

Why would a client pick you to deliver an application because of your sysadm skills for application server Random vX.Y? If you make your application correctly (possibly with a few work arounds), your application should deploy equally well on all three.

Weblogic and Websphere have one thing that JBoss will never get directly - a corporation that backs the product and promptly deals with support. And that is what you pay for.

Re:Jakarta Plug & My AppServer Experiences

[trilucid]

"What JBoss needs is a certification (with levels) for developers to obtain. If I go to a client and say "I have a level 3 WebLogic certification, a level 2 WebSphere certification, and know JBoss", what are they gonna pick?"

Hmm... makes one think, eh? As a developer, I've seen a lot of "certification wars" in the corporate contracting world. Here's my take.

The problem with "level-ified" certifications kinda resembles the "megahertz myth [to quote Apple]" issue. If you're assuming the client is a techno-yokel, you run into this problem with such cert programs.

Imagine, for example, two imaginary Linux certification programs. The first program (call it "EZLinux") sets out their certification map as follows:

• Level 1: Ability to use rm, ls, cp, and mv commands.
• Level 2: Understanding how to use RPM and DEB packages to update and modify a system.
• Level 3: Ability to use fdisk to create and manipulate partitions.
• Level 4: Actually got Mandrake running with help from the friendly neighbor kidz.

Okay, so that's an example of a *terribly* useless "certification" program that wouldn't be worth the paper the cert was printed on. Let's look at the other ficticious program, called "UR-Uber-H4x0r Linux":

• Level 1: Ability to quote verbatim the man pages for all Mandrake 8 standard linux commands (doesn't necessarily require deep understanding, just inhuman memory).
• Level 2: In-depth knowledge of kernel configuration and compilation, demonstrated by ability to correctly by hand [no Xconfig for you, for added flavor] compile a 2.4.x kernel for every known supported platform in existence.
• Level 3: Linux Torvalds willingly calls you Daddy, and calls you up for kernel hacking advice. Alan Cox routinely shows up at your pad asking for tree contributions.

Now, if you were Joe Hiring Manager, you might not actually know the difference between the two programs. Joe might look at them both, and say "wow, that first one has an extra level, so it's gotta be better!" ...

Companies will always try to use these tactics to make their products/programs/certs seem better than the others out there. Now, here's the real kicker: if Joe Hiring Manager actually understands why a certain cert is better than the others, he also (in all probability) understands why the product the cert is for is better. Hence, the better product wins. The key is education.

Just my take, that's all :).

Web hosting by geeks, for geeks. Starting at $4 USD per month.
If you're gonna email, use the public key!

Re:Jakarta Plug & My AppServer Experiences

[trilucid]

Geezus, I just called Linus Torvalds "Linux Torvalds". Damn this beer... :)

Web hosting by geeks, for geeks. Starting at $4 USD per month.
If you're gonna email, use the public key!

Re:Jakarta Plug & My AppServer Experiences

[FinnishFlash]

Weblogic and Websphere have one thing that JBoss will never get directly - a corporation that backs the product and promptly deals with support. And that is what you pay for.

Yeah, you pay for it allright ! That part is correct. Promptly deals with support...

Oh, please... You obviously haven't been working with IBM.

At least in Europe the IBM WebSphere "specialists" tend to know a great deal less of their appserver than ex. I as a developer. (there are exceptions of course.)

Their support is something like "Is the computer connected to the mains ? Oh, it is. Then we'll have to forward your request to the lab. They shoud be able to answer you in a week or two."

Somehow, I've allways had this feeling, that someday someone will point at IBM and say "But he has no clothes on !"...

Re:Jakarta Plug & My AppServer Experiences

[lqd]

Weblogic and Websphere have one thing that JBoss will never get directly - a corporation that backs the product and promptly deals with support. And what is what you pay for.

um, i've had the "pleasure" to deal with IBM WebSphere support in Europe. The guy on the phone said: you'll get the answer within 4 working days, we have to contact the guys in the US for this particular problem. i solved it 2 hours later by myself ... and it wasn't a problem with the support contract, I called with our customers support number and he has about 60 AS/400 boxen and various AIX machines sprinkled in between.

Project Info Page

[Foresto]

Jboss web site

Re:Project Info Page

[czardonic]

Score: -1, Obvious

Re:Project Info Page

[Anonymous Coward]

There's no need to be rude. The interview page wasn't loading for me at the time, and I didn't notice that its URL was that of the project itself.

What about running in production?

[Gollo]

Let's face it - JBoss appeals to us Slashdotters because

(a) It's open source
(b) It has a whole heap of fantastic development features.

What I didn't see an emphasis on is running on a daily basis in production. Sure, I think that JBoss is fantastic for development, and most of the leading edge features are great for developers, but what about running a mission critical production system? What benefits does it provide in that arena, given that if I have Weblogic or WebSphere, and it breaks on my 24x7 website, I can scream at the respective vendors?

Develop with JBoss, deploy with WebSphere/Weblogic. Anyone enlighten me to benefits of JBOSS in production over a commercial offering?

Gollo.

Re:What about running in production?

[Anonymous Coward]

If you aren't using J2EE, but just Servlets & JSP's, then I suggest Tomcat 4. They've really improved tomcat mucho (especially if you have to work on a Win32 machine). But only for development.

Another reason is that Tomcat & JBoss aren't resource hogs, when you need to be debugging and coding with it running).

Re:What about running in production?

[DarylBeattie]

Ah, I think, since you are unsure of the facts surrounding the reliability of different appservers (as am I), that you are really more concerned with "who to scream at" should your appserver go down.

One of the great fallacies of application hosting is that if there is somebody to scream at, you are somehow less responsible to your clients for the production screw-up than you would be if there was nobody but open-source developers to be frustrated with. [Yelling at open-source developers seldom helps your cause, no matter what it is.] Let's face it, your clients aren't going to care if you are blaming a commercial entity for the screw-up; it makes you look bad.

What you should be concerned with are:

1) Reliability - JBoss is more reliable than all other AppServers in use right now, and introduces cost savings because it is easier to use, and less buggy!

2) Support (NOT "blame") - I have used purchased appservers (well, admittedly "appserver"), and JBoss, and let me tell you; the JBoss group helped me quickly and easily with any problems I have had, wherewas the commercial product I was using was IMPOSSIBLE to get support for, even though my employers had paid big $$$ for it. [The same actually goes for all open source projects I have used.]

Daryl.

Re:What about running in production?

[jeet_means_win]

I think if you have a mission critical application then it is not right time to move to JBoss straight away, You should wait for 3.5 or 4 to be released. What Jboss really needs is backing from Big gaint like IBM etc.. this is what Marc, JBoss Founder/President, is trying to do create awareness, gain popularity. Once that is done application server market will be like HTTP server market... Who pays for Http server nowadays???

Re:What about running in production?

[Anonymous Coward]

I can see you have been reading Big Blue's and BEA's marketing literature by your use of "mission-critical", "production", and "24x7" in the same paragraph.

In my experience, all screaming at BEA is directed at the nearest brick wall they have handy. The most difficult thing about JBoss is the relative scarcity of coherent, directed documentation in a usable form. This makes one-time configuration take somewhat longer. However, with the addition of clustered persistence and failover capabilities to JBoss (3.0), it really competes nicely with the other two. The big difference is price-point. JBoss is free.

Moreover, despite what the marketroids claim, WebLogic and WebSphere are both monolithic architectures, which means that they lag the evolving standard by a ways. JBoss tends to be more current.

Re:What about running in production?

[glwtta]

Have you ever actually tried screaming at BEA? It's one of the most frustrating and fruitless endeauvers you can find yourself in. I can't speak for IBM, but in BEA's case, screaming at them is useless. Couple that with their shitty JDBC implementation, and you've got yourself days of fun for the entire family.

(I know some people said this already, but I figured this is something where every vote counts.)

Re:What about running in production?

[christophersaul]

Companies who want something that's faster than Apache and that's supported by Professional Services teams who can design and implement a solution, hence the fact that iPlanet and other web servers are still very much with us.

Re:What about running in production?

[StandardDeviant]

I seriously doubt that, unless IBM is astoundingly stupid, that they will advocate JBoss unless and until WebSphere (you know, that thing their banner ads on slashdot are always talking about, their $$$$$$$$$ java thingie) is not being sold anymore. IBM doesn't make any bucks off of AIX really (compared to the bucks they make on the hw anyway; c.f. solaris and sun), so they have no problems "replacing" it with Linux. Now, I can very much see Sun doing this, as they already support Tomcat as the official reference implementation of Servlet API foo and JSP API bar, and further they don't sell a java app server product that I'm aware of (if they do have one they're not exactly trumpeting it). [If sun started pushing JBoss with Tomcat, I imagine it wouldn't be much of a leap to advocate PostgreSQL as part of that solution. Yes, I am a hopelessly rabid Pg fan. ;-)]

IBM is a hardware and consulting company with occasional delusions of being a sw company, so sometimes they will have incentives to push open source and sometimes they will not (e.g. notice that they don't support any open source database projects, hmmm, db2?). Sun OTOH is much more of a pure-play hw company (which is why their stock is pretty hammered right now IMHO), so they have a greater incentive to push open source software.

Think middleware

[Anonymous Coward]

In general, application servers handle most of the issues regarding connecting and integrating groups of applications and exposing them as a collective entity.

So for example, the application server may handle pooling of database connections from other code. Also, the application server may be responsible for monitoring other code and altering resources as required. Clustering is an obvious next step for these servers, with load balancing on top of that where applicable.

Overkill for anything that's not mission-critical?

[Anonymous Coward]

I wonder -- wouldn't someone developing a middle-sized website, or even a non-mission-critical one with the enormous traffic of Slashdot, be better off using the tried and true "Linux Apache MySQL Perl" (or adequate substitutes) combination? Isn't JBoss only for the kind of application where you need sophisticated exception handling, load balancing and clustering?

Evangelist? Say it ain't so!

[tswinzig]

was working at SUN on Java since the early days, mostly as an evangelist in the beginning.

After reading his responses, you can really tell he had a job as an evangelist. For you Quake/id fans out there, it's like Paul Steed started coding in Java!

Re:Evangelist? Say it ain't so!

[Anonymous Coward]

i dont get it

hmmmm

[Anonymous Coward]

Damn.

I guess I must live under a rock. I thought JBoss was some "urban" clothing conglomerate, formed when J. Crew and Hugo Boss merged to oust Tommy in the Phat Ghetto Style Stankfinga razmadaza hootchie combo platta market.

Y'knowumsay'n? Yomesain?

Server-side Diversity?

[Anonymous Coward]

> I was working at SUN on Java since the early
> days, mostly as an evangelist in the beginning.

Hmmmm...

> I grew rapidly convinced that Java's main
> feature, its portability across Operating
> Systems, was not really worth having on the
> client-side

"Oh no, the Java UI isn't getting any traction! Quick, spin with some lame excuse!"

> but extremely relevant on the server-side. The
> reason being the diversity of OSes and hardware
> platforms on the server side that does not
> exist on the client side.

1. Why mix platforms on the server-side? It doesn't make any sense.

2. Java makes lots of sense on the client side, where your box might be Windows/Mac/Linux/Solaris...too bad the UI is horrible.

Re:Server-side Diversity?

[dildofire]

exactly. nobody wants to mix platforms on the server side. companies will usually pick one architecture and stick with it (which makes more sense, less types of hardware to support). and besides, most modern programming languages are portable anyway. java's big deal is that you don't have to recompile to change platforms, but that only takes like 15 minutes anyway! the bottom line is, if you write your code correctly, portability shouldn't be an issue with most modern languages.

Re:Server-side Diversity?

[Space cowboy]

Mixed platforms on the server side are the norm rather than the exception, in the market where JBoss is trying to position itself.

Consider that JBoss is (will be, in some respects) a J2EE-compliant system, in everything except name. The purpose of J2EE is to deploy in large-scale environments... where you potentially have PC/W2K, Solaris, IBM S390, and God knows what else already providing a business solution.

You will not get a contract to replace the business systems from top to bottom, so take the view that you'll integrate into them instead. Now you're talking multi-tier Facades and Interfaces in order to get these things talking to each other.

You'll want to design things so that there is as little interdependence on co-operating subsystems as possible. Read 'Design Patterns'. Have a serious *think* about XML. Look at the Java libraries and the J2EE designs. Understand.

Have you ever noticed that it "just seems easier" to do things in Java than C/C++/VB(spit!) when there's any degree of complexity ? (I'm not talking about a couple of days' hacking, I'm talking about a designed system). There's a good reason why Java works so well - it's been designed well from the bottom up.

J2EE (and hence JBoss) take that elegance to a higher level of abstraction, but what they're doing is to continue the excellent design principles of the class libraries.

A happy JBoss user.

Simon.

Re:Server-side Diversity?

[Anonymous Coward]

> The purpose of J2EE is to deploy in large-scale
> environments... where you potentially have
> PC/W2K, Solaris, IBM S390, and God knows what
> else already providing a business solution.

Gotcha. I can see people loving something like JBoss in that situation...and am happy I have one Linux platform to run on. :)

> Have you ever noticed that it "just seems
> easier" to do things in Java than C/C++/VB
> (spit!) when there's any degree of complexity ?

I don't know if I agree there (other than VB). I think there are open-source/free C++ alternatives to the Java framework that rival it in ease of use and power. I find that the more I learn about C++, the more elegant design patterns I can come up with.

Happy C++ hacker. :)

Re:Server-side Diversity?

[Second_Derivative]

...I think there are open-source/free C++ alternatives to the Java framework that rival it in ease of use and power...

Care to elucidate me? =)

Re:Server-side Diversity?

[e2d2]

Look at it from a developers view point. If I know Java and use Jboss I can develop server solutions for Both Windows and *nix platforms. It s pretty easy using Jboss or similiar app servers. Even if it was true that enterprise organizations use all the same hardware/platforms (laugh), what about the outside developers that create apps for them? It makes more sense (read it makes more money) to develop in Java in these scenarios.

Re:Server-side Diversity?

[Anonymous Coward]

15 minutes to recompile??? Have you never worked on a real project? Maybe if you're just compiling a couple parts, but if you want to do a serious system that's got 5 years of work behind it, you're looking at a few hours per platform.

Honestly, the compile time is meaningless anyway. It's the necessity of testing and debugging on multiple architectures that sucks your time. With Java, you have few to no extra tests to run (if you did things the Right Way(tm)).

Re:Server-side Diversity?

[fredrik70]

"I find that the more I learn about C++, the more elegant design patterns I can come up with. "

yes, However, it's very easy to blow your foot off (or your leg off!) with C++. One has to be careful when using it, having said that, that doesn't mean you can just sit down and hack away in java and come up with a beautifully designed system! ;-) C++ does require more though IMHO though, since it provides you with quite a bit more powerful tools (templates, operator overloading, multiple inheritance), which are both a blessing and a curse, depending on what day it is ;-)and don't provide you with 'java's little helper' (Garbage collection, and, eh, other stuff - i dunno about)

Downloads as market share??

[Anonymous Coward]

JBoss is the hidden variable here that none of these guys want to talk about. Since we give away our product, the true statistic to look at is number of downloads, which we find to be a more honest way of measuring market penetration.

Come on, does he even expect that 10% of the people who download it, actually run it????

Re:Downloads as market share??

[Anonymous Coward]

No, but considering the leading server from BEA gets a download number of couple of thousand a month for their evaluation version, that is still a very very impressive number of downloads.

Open Source J2EE Stack

[adamy]

At my office we use JBoss as part of an entirely Open Source and free platform stack:

OS: Linux
Web Server: Apache
Servlet engine: Tomcat
EJB Container: JBoss
DB: PostgresQL

And a great set of middle level libraries such as Struts for form processing, a slew of other jakarta classes, tinySQL fro xBase integration, JUnit, and HttpUnit.

I came from a company that did oracle/ATG integrations. I can honsetly say as a developer, I have everything I need that I used on thos platforms.Plust I like that fact that it is a little closer to the J2EE standard than ATG.

Re:Open Source J2EE Stack

[MSBob]

Wow. Where do I apply :). In fairnass though it's nice to see that OSS stuff is already reaching the point where you can put together a very decent enterprise app purely out of OSS components. I betcha the performance is ten times better than if you went with a hog like WebSphere.

Good Stuff.

Re:Open Source J2EE Stack

[gupta]

You are blessed. i will have to work for evil empire soon. that is my only option now.

Java Application Server

[peterdaly]

A java apllication server is a software package that runs the logic of many high end java applications. An application server is where most of the hard work of an application is done. On top of application servers, you may see things like Java Applets, Java Applications, or JSP/Servet containers (ala Tomcat) which act as the interface between the application and the user.

Usually on a Java Application Server, programs are design as small server side components, which perform independently of each other. By typing the components together, useful applications are born. For instance in e-commerce, a JSP interface may recieve a credit card purchase. It will run your credit card through a credit card component on the JAS, ask a warehouse component to mark the order for picking, notify the shipping component a package will be coming and where to ship the order number to, notify marketing a sale has been made, and tell the purchasing module to order parts which will be needed to replace the unit in inventory.

All of these are seperate components which can be used in many different applications. By creating a system like this, business login never has to exist in more than one place, which reduces programming time, stabibility, and makes the system as a whole more flexible.

Hope that made sense.

-Pete

Re:Java Application Server

[Anonymous Coward]

Ah now that's a good explenation. What I wonder is will this be running physically on one machine ? Wouldn't it make sense to put certain parts behind a firewall or on other machines ?

Re:Java Application Server

[easter1916]

Yes. At the large site I currently work on (20,000+ concurrent users) we run the web tier (JSP and servlets) on clustered Sun servers, each running multiple server instances, in the DMZ and the business logic and database stuff (EJBs on a similar server setup behind the second firewall.

My JBoss experiences

[pHalec]

I've been using JBoss for a little over a year, and my experience has been very good. It's fast, reliable, and seems to keep quite up-to-date with developing standards.

This kind of project is exactly what the free software world needs - usable, cutting-edge, and easily comparable with competing proprietary products.

This project has cemented my opinion on Java on Linux as a server platform, and when combined with PostgreSQL, it forms a complete and surprisingly robust setup.

I wish the JBoss team the best of luck and fully intend to keep using and recommending their software.

Not whoring but...

[Fnkmaster]

First, it seems like a lot of posters don't know what an "Application Server" in Java parlance is. JBoss and the other mentioned products in the post are really implementations of the J2EE or Java 2 Enterprise Edition API set, as specified by Sun. This is not in any way opposed to J2SE (Standard Edition), which is the baseline Java SDK (VM + API library), but is an addition to that - it includes the EJB (Enterprise Java Bean) API for writing transactionally aware business logic components and connecting them to persistent data sources (usually RDBMSen of various sorts). I'm mixed on EJB, having built several large systems with it - it's just too damned easy to make applications that suck with EJB if you don't know how to use it. The CORBA CCM spec seems a bit better, with some kinds of components that I consider very useful that EJB doesn't support. Oh well. J2EE also envelops JNDI, a very useful and underutilized naming and directory API that solves lots of VERY common problems in building enterprise apps (the old chicken-and-egg resource location/distributed configuration problem can be solved very easily with JNDI though few people correctly use it). JMS is a decent enterprise messaging API, essentially an API for reliable asynchronous communication between distributed components either within or between enterprises that abstracts the details of writing socket level communications, then worrying about bridging firewalls etc. from the application developer. But it has some serious flaws in the way it forces you to do things. Nevertheless people have addressed these in not-fully-compliant add-ons to the spec to make it more like successful, established products such as TIB Rendezvous. (there are other parts of the J2EE spec, like JMX and the Servlet API, and probably a few others I'm forgetting right now).

Now that J2EE has been briefly explained, and I've stated my position (useful services, with some warts and all, designed to the lowest common denominator, and unfortunately sometimes too easy to build systems that really perform like slugs) I will give JBoss some props for really driving forward the implementations of the standard in several ways. Adoption of JMX is great, JMX is very useful for building custom manageable components that don't fit into the standard J2EE framework (i.e. they need to be stateful and not session coupled so they can't be implemented as reasonably performing, compliant EJBs). Also JBoss provided the first reasonably performing EJB implementation I have seen. Far faster than most of the commercial implementations when it came out for the common case scenarios (the commercial implementations may have improved since, I don't really know). My company moved to JBoss from Weblogic as we discovered we couldn't afford enough Weblogic licenses for every developer to have his own test box. And that the nature of our system makes it really hard to test otherwise (note: this is partially the fault of our system's architectural stupidities, but let's put that aside for the moment). We have generally had great luck with JBoss, and found the JBoss community to be very, very helpful with problems when they sprung up (compared to Weblogic 5.1 where I got some very frustrated engineers on my team stuck with the job of calling Weblogic support bitching at them about weird problems with spontaneous breaking of the EJB standard - anyone who used it knows that SP6, SP7 and SP8 were all released in short succession around the time I'm speaking of).

The moral of all this is that the JBoss team has done a fabulous job at providing a great, useful product that has saved my company thousands of dollars and many hundreds of man-hours of developer time. While the J2EE spec is, err, deficient in certain ways, for a lot of enterprise software projects it's good enough for the task at hand. And JBoss, with full clustering support coming up now, should be good enough for most if not all of these jobs. If you need a real distributed application with high volume transaction processing, you might need to look at other kinds of systems that give you more access to lower level capabilities (think: Weblogic Enterprise, Tuxedo, etc.). Or roll your own. :)

Another server

[Xunker]

Another server that is used in the Java arena is the Orion Server. It's very nice and I enjoy working with it on a daily basis, but it's not Open Source which a lot of people consider to be a downside. It's free for development platform and non-profits, but for production it's $1500 USD per host. Cheaper than BEA, but But a lot more expensive than Jboss or TomCat (the Apache JAS).

Re:Another server

[Fnkmaster]

Orion is a fabulous product - its weakness used to be terrible documentation (the last time I did a thorough J2EE server overview). I mean, really weak. I couldn't figure out how to do even relatively basic stuff - a lot of their documentation appeared to be written by non-native english speakers. Some of this applies to JBoss too, but JBoss has enough documentation to get you "close enough" for most things you need to do, and when that isn't enough it's got a great community, user forums, old mailing list archives, etc. which have proven very useful to me for cases where the obvious stuff didn't work and the documentation was poor or out of date.

Re:Another server

[perrin_harkins]

The Orion server is so nice that Oracle bought it and now sells it as their J2EE server.

Re:Another server

[farkas]

You should take look at the new documentation that we have put together. http://kb.atlassian.com.

Indexed, searchable, cross-referenced & annotatable knowledge base for Orion information.

Re:Another server

[nettdata]

It should be noted that the Orion Server has been licensed by Oracle as the main J2EE component of their 9i Application Server, and does a great job in both performance and fault-tolerance.

Re:Another server

[stardeveloper]

We are using Orion Server on our web site ( Stardeveloper.com ) for about six months now and one thing I must say is that Orion is excellent. We have not yet faced any problems with it. It runs like a breeze. Oh! and not to forget, we are running Orion on Windows platform with Sun's JDK 1.3.1 and haven't had to reboot the system in months.

Orion takes so less memory ( 35MB out of 512MB available ) and runs so fast, I would suggest everyone to at least give it a try, and if you not thrilled with it's performance, go Tomcat/JBoss way.

I have done all kinds of performance tests on Orion, Resin, IIS, .NET Beta 2, and Tomcat. By far the fastest is Orion, then comes Resin ( which doesn't support EJBs ).

P.S. One thing lacking in Orion is that it currently doesn't fully support J2EE 1.3 specification. Although it does support CMP 2.0 beans and Message-Driven beans, it's support for Local interfaces in EJBs is lacking.

J2EE Rules

[Anonymous Coward]

everything I know about J2EE I learned from peterdaly

Java - the perverse language

[pete-classic]

Okay, Java code itself is pretty easy to deal with. Pretty much C++ with a few differences (final vs. const, etc.)

But geezus if they don't make up terminology just for the hell of it. I've tried and tried to get my head around a little of it. Not much luck. I must have looked at fifty block diagrams showing how the JDK, the SDK, the JRE and freaking EJB fit together, and it still makes no sense.

What am I missing here? I understand compiled languages and interpreted languages/scripts. I do a little assembly. I understand the overarching concept of the VM and "bytecode" (which, best I can tell is just machine code for an imaginary machine called "JVM") but it took me weeks of trying just to get to the point of "hello world." Do I need the JRE and the JDK? What is the difference between the SDK and the JDK? Okay, I found the download page of the JDK, but when I agree to the license it just refreshes the page! Holy shit.

And let's not even start on variable names like "The_Longest_Yet_Least_Descriptive_Method_Name_I n_The_World." And don't dare drop the capital "t" in "the" in your call, or you'll get an error message in Sanskrit pointing to nine lines before your call. Oh, and you know what else is a great idea? Make everything sensitive to the freaking filenames of the source files.

Is there some mind-set thing that I am just missing? Is there some parallel-universe where this sort of stuff makes sense?

Frankly, the whole thing strikes me as being like Psychology (and a lot of semi-sciences) where perverse vocabulary is used as a barrier to entry to the field. (Oops, too many "lay people" know what MPD is! Quick, change the name to DPD!)

-Peter

Re:Java - the perverse language

[Anonymous Coward]

Read "Thinking in Java" by Bruce Eckels, and it will put your mind at ease.

Your first experience, its sounds, like you tried to swallow the whole thing at once, and choked.
-or-
You were biased to Java and subconsciencly didn't want to learn it
-or-
You hardly know Java, and are trolling (and I bit).

Re:Java - the perverse language

[jeffy124]

difference between JDk and SDK: Prior to Java 1.1 (inclusive) the developer kits were called "Java Developer Kit" (or JDK). From Java 1.2 on, it's been called the "Software Development Kit."

The SDK contains what you need to compile and run programs. The JRE (Java RunTime Environment) is just the run part of the SDK. The JRE comes with the JDK.

The concept of having source files match the class name within is required for classes that are 'public.' It's also good development practice to do it that way.

As for the download problem - check your browser settings. You might need javascript/cookies or something.

Dont worry too much about the over-acronyming and over-buzzwording of all the Java stuff out there. It's nearly impossible to keep it all straight. (for example, I dont bother with about 95% of the java stuff out there) But to learn what you need just takes the usual: practice and experience.

Re:Java - the perverse language

[iapetus]

I disagree with you on virtually every front here: Java is far from perverse, and a lot of things you're criticising it for make sense if you think about them. Let's take it one point at a time:

Do I need the JRE and the JDK?

No. The JRE (Java Runtime Environment) is for running Java applications only. The JSDK (Java Software Developers Kit), also known as the JDK or occasionally by the more generic term SDK, provides the same functionality but also the ability to compile Java code.

And let's not even start on variable names like "The_Longest_Yet_Least_Descriptive_Method_Name_I n_The_World."

Let's not, unless we can give examples. Most Java variable names in Sun's own code are fairly self explanatory once you know the basic structure they're working in.

And don't dare drop the capital "t" in "the" in your call, or you'll get an error message in Sanskrit pointing to nine lines before your call. Oh, and you know what else is a great idea? Make everything sensitive to the freaking filenames of the source files.

Actually, in that case the compilation error would be on the line where the error occurred, as with most error reporting in Java programs.

As for case sensitivity, it's pretty much required for a platform-independent system that's going to be running on Unix-like systems.

(If you want to be really frightened, be aware that Java source code is actually UTF rather than ASCII, so you could create a file where all variables are called 'bob', just in different character sets...)

Re:Java - the perverse language

[pete-classic]

Well, I expected to be modded down, but I am sincere, so I guess it depends on what, exactly, you mean by trolling. I do "hardly know Java" but hardly for lack of trying.

As to "swallow[ing] the whole thing at once." I did, in fact, expect to be able to get to "hello world" within the first day or two. Since I had gotten at least this far with several other languages I didn't think that this was unreasonable.

I am consciously biased against Java, but I am pretty sure that I'm not so sick that I burned hours of my life to prove to myself that I don't like it. I was trying to learn it for my own amusement, and would still like to.

Bruce's site (with the flash intro with the background that almost matches the background of the rest of the site) has, sadly, made my mental state worse. I hope that you, AC, aren't Bruce. (I suspect you aren't, since you can't spell subconsciously :-P )

[time passes . . .]

I'm back from the site, and I have to say that Bruce kicks ass. I downloaded the Java book (and the C++ books) and am looking forward to curling up with my laptop and checking them out.

-Peter

Re:Java - the perverse language

[pete-classic]

No. The JRE (Java Runtime Environment) is for running Java applications only. The JSDK (Java Software Developers Kit), also known as the JDK or occasionally by the more generic term SDK, provides the same functionality but also the ability to compile Java code.

I understand this, now. But (at the time) there was no good way to figure this out form Sun's Java site. Now there are actually some pretty good walk-throughs.

Let's not, unless we can give examples.

Well, I don't have any Java code handy, but the stuff my wife has had to use for her Java classes (as in school) has been pretty bad. Weird capitalization, way to long without being particularly descriptive, etc.

I think we can at least agree that typical method and variable names in typical Java programs are longer (and therefore more subject to typos) than most programs. It is my opinion that in the general case there is little (if any) useful information embedded in this baggage.

Actually, here is an example from "The Practice of Programming" (Kernighan & Pike):

? class UserQueue {
? int noOfItemsInQ, frontOfTheQueue, queue Capacity;
? public int noOfUsersInQueue() {...}
? }

[comments on style]

class UserQueue {
int nitems, front, capacity;
public int nusers() {...}
}

since it leads to statments like

queue.capicty++;
n = queue.nusers();

No clarity is lost.

This quote looks pretty bad. I have the trade paperback version, and this appears on page 4.

Anyway, to my eye, the first looks distincly "Javaish" whereas the second looks like C++. I'm not sure, but I think that this code is valid in either language. That's my point about perverse, typo prone method and variable names.

I fully expect case sensitivity. I don't, however, expect 20+ character labels with "funny" caps.

Actually, in that case the compilation error would be on the line where the error occurred, as with most error reporting in Java programs.

Well, of course this isn't true. It isn't true for any language. Any compiler gives an error where the compilation error is encountered. So, if you forget to uncomment a declaration you get the error where the variable is used.

OTOH, I get the sense (after debuging my own bad syntax on C++ and Perl) that "javac" (ask me how long it took me to figure that out) encounters errors in places less inutitive, and further from the mistake than others. Again, this is a sense, I don't have examples handy.

Anyway, this is pretty far off of my core complaint, which is that Java seems to be more about buzwords and barriers to entry than about technology.

I do have one more question: is there an assembler for the JVM?

Okay, one more: is there some documentation for the JVM in terms of things like registers, service calls, etc.

-Peter

Re:Java - the perverse language

[Algan]

A couple of suggestions:
Get a GUI IDE, I'd suggest JBuilder, it does syntax checking on the fly while you type. I never had a syntax error at compilation after I got it. You can download the personal edition from Borland for free. It's written in Java, so it will run happily on Linux
Download the Java tutorial from javasoft.com. It's the best tutorial available. It's also available in print, however the printed edition tends to lag a bit.
Download the full SDK documentation and keep it for reference. You're going to need it.
The entry barrier is low in my opinion, at least for somebody who's familiar with C++. Once you start to grok it, you'll like it.
OH, and there are several documents describing the JVM (which if I remember correctly is a Stack machine, so it doesn;t have registers). Just browse javasoft.com

Re:Java - the perverse language

[iapetus]

I think we can at least agree that typical method and variable names in typical Java programs are longer (and therefore more subject to typos) than most programs. It is my opinion that in the general case there is little (if any) useful information embedded in this baggage.

That's the general style, yes: meaningful method and variable names. If you prefer to take other approaches in your own code then there's nothing stopping you from doing so.

I don't, however, expect 20+ character labels with "funny" caps.

You're confusing 'funny' caps with a well-defined style for capitalisation. Java variable names, if you're following Sun's style guide, begin with a lower case letter, and new words are identified with a capital letter, an approach sometimes referred to as IniCaps. Method names use the same style, class names the same but with an initial capital letter. Once you're used to that style, it makes perfect sense, and you're a lot less likely to drop caps. It's preferable to some of the overly abbreviated variable names you get in some C code as far as I'm concerned - particularly for maintenance purposes. I know what the variable userCount contains, or can make a good guess. I have no idea what nusr means.

It's all down to style, of course. There's no reason a C program couldn't be the one with a variable userCount and a Java program the one with nusr, and I've seen both.

Anyway, this is pretty far off of my core complaint, which is that Java seems to be more about buzwords and barriers to entry than about technology.

It's true of all languages that until you know the jargon, you will find it confusing. That's the case with Java to the same extent as other languages.

I do have one more question: is there an assembler for the JVM?

A somewhat unhelpful answer here: yes, but I don't know where you'd go to find it.

Okay, one more: is there some documentation for the JVM in terms of things like registers, service calls, etc.

I guess this is what you want.

Re:Java - the perverse language

[pete-classic]

The IDE advice sounds like an admission of guilt to me. I'm not saying that a piece of software that watches you for mistakes in inherently bad, but a language that makes it that important is.

Wow, no registers. I'm not saying you are wrong, but I can't imagine how this can be . . . how do you keep track of the top of the stack!

[read, read, read]

Oh, there is a "pc" register which seems to be equivalent to IP on intel. I can't find anything on a stack pointer register. Weird.

-Peter

Re:Java - the perverse language

[ameoba]

Uhh... Hrmm, "hello world" in Java..

import javax.swing.*;

public class HWorld
{
public static void main(String args[])
{
JOptionPane.showMessageDialog(null, "Hello World");
// alternately, you can print to console
// System.out.println("Hello World");
// either of which makes a lot more sense than
// C++'s "cout

Re:Java - the perverse language

[pete-classic]

Thanks for that link. I found that via the general link provided by the other guy that replied.

I find the name "IniCaps" to describe this laugh out loud funny (since it doesn't comply with itself).

As to variable names, you seem to have ignored two key points in my post. First, in my opinion, the longer names don't actually carry additional info in practice. If you are just saying that in your experience they do, well, that's fine then.

The other is that, as illustrated by the example, the full name (I'm sure there is some cool computer science name for this) can be made to look very good, and it doesn't matter if the name looks vague out of context. This isn't my point, it is K&P's point.

It's true of all languages that until you know the jargon, you will find it confusing. That's the case with Java to the same extent as other languages.

You must be trying to miss my point here. Yes, of course every technical field has its jargon. And every language its own set of jargon. My point is that Java (I guess really Sun) goes way, way beyond what is normal.

One prime example: EJB. You seem to have a personal affinity for meaningful labels. Let's expand EJB. AFAIK it is Enterprise Java Beans. To me that is a buzzword followed by a (maybe) cute but meaningless bit of wordplay. NO INFORMATION. I heard a rumor this is an API. IMO this is going as far as to be deliberately evasive in naming a technology.

-Peter

Re:Java - the perverse language

[ameoba]

please disregard... the left-shift operator got eaten by the HTML parser, and I lost the 2nd half of the post.. +)

See... Java's simpler than HTML. +)

Re:Java - the perverse language

[bill_kress]

It's funny how often I've come across C++ programmers who said things like this before they used Java.

After using java for a month or so, it's really fun to hear them piss and moan about having to go back to C/C++

Generally, the first thing that starts to convince the hardcore C/C++ hackers that java isn't just a glorified C++ Clone is when they start to use the Reflection package.

The libraries can be a bit daunting at first, I can understand why you were overwhelmed, But if you start here:

http://java.sun.com/j2se/1.3/docs/api/index.html

and study one package at a time (in order of interest) it really isn't bad. It's just that most of those things in C/C++ that you'd have to buy or create externally are already built into Java's default class library.

This is actually a huge advantage, by the way, since all the packages are integrated with each other and operate consistantly.

I used C/C++ for quite a few years before moving to Java, and after the first few weeks of Java I never looked back. I'm working for an all-java company now, and couldn't imagine moving back to C++ (I'd use C if it were necessary).

Now, by the way. You are evaluating how "Java Programmers" use names by something from your wife's class. Do you really think this is fair? Our company policy is the shortest descriptive name you can come up with... period.

Also, javac has become AMAZINGLY adept at telling you exactally where an error is, and is right so often that when it isn't I'm completely shocked. In fact, the only time I can remember that this happened, I was using one of the open-source java compilers, not javac.

Re:Java - the perverse language

[FatHogByTheAss]

So, your basic complaint seems to be "I don't know Java, I'm too dumb to figure out Java, my old lady took some Java in school, and she's too dumb to figure it out, too, so it must be jargon."

Here's the deal, scooter. EJB is a real technology, with it's own specification. The simple fact that you "cant get your head around it" speaks far more about you than it does about Java.

So you go on and continue to whine about your misconceptions. The rest of us are going to continue to write beans, run the cool containers, and deploy them.

Love!

FatHog

P.S. Since you seem to blame your tools for your own ineptitude, here is a clue for you.

Name this language:

$FatHogHitMeOverTheHeadWithTheStupidStick = 1;
%IShouldGoBackToAccountingWhereIBelong = qw( Peter => 1, FatHog => 0);

bless my $soul, 'Silly';

Re:Java - the perverse language

[Anonymous Coward]

Java's sell to it's corporate masters is that it easily maintainable by (cheaper|easier to find) programmers.

I think we've all see too many cases where nusrs was "newUsers" and not "numberOfUsers" (and so on). Bug waiting to happen.

Re:Java - the perverse language

[pete-classic]

Actually, the complaint about EJB had nothing to do with the technology. I can't criticize it, because I don't know anything about it. Except that it has a non-descriptive name. If it has to do with "containers" why not call it JavaCan (or would that be javaCan?) or numberTenCan (NTC, sounds nice) or whatever?

And my complaint isn't that I am too dumb to learn Java, it is that I seem to be having a significantly higher level of difficulty getting started with Java than I have with other languages in the past. (Of course, I grouse almost as loudly about Perl being too easy in some ways . . .) So, maybe I'm dumber than dirt, but that doesn't appear to be the only issue involved.

Actually, the things I have learned in the course of this thread lead me to believe that the "problem" (maybe it is aesthetic, maybe it is just me, whatever) is due to over-marketing and over-committeizing. IOW, I'm not sure any sane person would arrive at "IniCaps" but a committee of sane people might end up converging on it. No language designer would come up with Enterprise Java Beans, but marketing department sure would. I guess the fact that it doesn't have a "father" (like Wall or K&R) shows.

Maybe we mean different things by jargon? Your message could be read as you are under the impression that since I call EJB "jargon" I don't think it is a real technology. Is that what you meant? Anything technical has jargon. It has to. I'm just saying that I find Java (in the larger sense of Java) to be jargonistic to the extreme, and for its own sake.

-Peter

PS: It is, of course, the only slightly less perverse language, Perl.

Re:Java - the perverse language

[iapetus]

JavaBeans is the component model for Java. Java - coffee. Beans - coffee beans. Isn't it cute?

Enterprise JavaBeans is the *enterprise* component model for Java.

There, that was fairly painless, yes?

As for your comments on IniCaps not conforming to itself, it's used to describe the process of capitalising new words (short for initial capitals) - both the variable and class name conventions can be described as IniCaps.

Re:Java - the perverse language

[pete-classic]

Yeah, except that the problem was that I couldn't figure out what PACKAGE I needed. (JDK, SDK, JRE, LMNOP, etc) I mean, I'm pretty competent, but this was kept like some state secret.

None of the human beings that I knew at the time seemed to get this either, and all the "cool guy" Java guys and docs just assume you know this.

As I said before, there are some pretty useful walkthroughs for first-timers now.

But since you brought up how much more sense System.out.println() makes vs. cout . . .

Since Java is all logical and stuff, I can get input by just doing System.in.getln() right?

Wah-wah-wah-waaaaah*

Anyway, how do you do the equivalent of "cin >> some_string;" in Java?

-Peter

* the sound on "Let's Make a Deal" when someone gets the donkey

Re:Java - the perverse language

[AndersDahlberg]

something like: String nameOfStupidPerson = null; try { BufferedReader br = new BufferedReader(System.in); nameOfStupidPerson = br.readLine(); } catch (IOException ioe) { System.err.println("A system problem: " + ioe.getMessage()); System.exit(1); } System.out.println(nameOfStupidPerson + " is a dumb ass"); Perfectly safe and a lot clearer than the c++ version... (and for heavens sake: don't forget checking forexceptions - or else your customers will kill you)

Re:Java - the perverse language

[pete-classic]

And for heavens sake: don't forget to preview - or else the trolls will kill you.

It looks really nice when I do "view source" though.

Anyway, thanks!

So the key here is "new BufferedReader". Am I correct in the belief that there is no "free br" allowed/needed?

-Peter

Re:Java - the perverse language

[pete-classic]

You are clearly illustrating my point.

Yes, java beans is cute, and I get it, but it doesn't tell me squat about the technology.

What does enterprise mean? I'm keeping a running list:

Supports files >2gb
Supports RAM >4gb
Supports fibre channel
We charge you really, really a lot more because you are a big company and we think you can afford it.
Supports LDAP
Has even more proprietary extensions than the "workgroup" version.

I'm not sure if any of these fit.

I think that the IniCaps thing might be even funnier now. Maybe it should be iniCapsCptIniWrd (Initial Capitals exCept Initial Word)!

Okay, I'm just being silly now.

-Peter

Re:Java - the perverse language

[Anonymous Coward]

The only language I've seen that requires an IDE is Visual Basic. I don't just mean it requires an IDE because that's the only way you can get it, I mean you'd never get a decent amount of working code unless you could memorize all the inconsistencies and general gunk that's required to make it work.

[public Rant]

I hate properties. Not only are the names often misleading, you have no clue how to use them. How am I supposed to know a propery can only be set to a long, not a string or a boolean, without looking into the docs all the time? Especially since it looks and acts like one thing, but is in fact quite another. Since you're forced to use an IDE, you could have nice methods which tell you what type to use, but no. Same goes for reading properties.

Don't get me started on functions. Do you have any idea how annoying it is to not have a return statement? It makes code more difficult to read because there's no definite exit point. The Exit Sub statment just doesn't stand out in a bunch of code. Beside that, if my function is named getSubMasterAdName, I don't see why I need to keep reminding myself of the name of my own function every time I add something to the end of the return string. Not to mention the fact that a simple name change for a function becomes impossible without examining the details of how it works to make sure nothing is missed.

Everything is just so quirky. You have to take notes on every single object you use just because most all of them work quite differently, even if they have similarly named attributes. With Java, I almost always have get/set methods, except for some old code (and perhaps new code that isn't well thought out) so it's usually not a hassle. Now all I have to do is figure out these stinking layout managers, and life without VB will be reality (unless my boss figures out that I stopped using it and gets mad or something).

And who says this:

Dim str as String
'...later, in initialization code
str = "Initialized value"
is better than this:

String str = "Initialized value";
???

[end Rant]

Re:Java - the perverse language

[Anonymous Coward]

I agree. javac is very good, almost all of the time. I still dabble in C, but I'll take Java over C++ any day.

Re:Java - the perverse language

[Anonymous Coward]

It's called the API docs. Honestly, just start with java.lang. If you can figure out how to use the API docs, you can do anything (almost ;).

Re:Java - the perverse language

[Algan]

Don't get me wrong, you don't HAVE to use an IDE. The command line compiler is quite good at reporting errors correctly, better than gcc IMHO. If you're a hardcore command line fan, then you can live happily with it. But personally I like to spend my time hunting bugs in logic not syntax errors, that's why I prefer JBuilder.

Re:Java - the perverse language

[FatHogByTheAss]

Actually, the complaint about EJB had nothing to do with the technology. I can't criticize it, because I don't know anything about it. Except that it has a non-descriptive name.

I'd like you to take a step back and read that. You know nothing about it. How can you tell me the name is non descriptive?

If it has to do with "containers" why not call it JavaCan

Because if you knew anything (and we've already established that you don't), you would know that EJB is a pretty descriptive name. It has three terms, "Enterprise," which means something in the context of systems, "Java," which also means something, and "Beans" which in the context of Java, also means something quite specific.

I guess the fact that it doesn't have a "father" (like Wall or K&R) shows.

Well, scoot, James Gosling and Bill Joy are pretty accomplished engineers. I'm sure they would be quite distraught that a little high school piker like you thinks they're a "committe."

I'm just saying that I find Java (in the larger sense of Java) to be jargonistic to the extreme, and for its own sake.

So long as you continue to wallow in your own ignorance, this isn't going to change. You have already acknowledged you know nothing of the technology. You have already acknowledged that you don't understand language style issues. Why do you continue to run on at the lip?

Re:Java - the perverse language

[AndersDahlberg]

ahh, that was what I was saying - "do not forget handling exceptions 'and closing io-streams ;)" hehe here you go:
BufferedReader br = ...
try {
... } catch (IOException ioe) {
...
} finally {
try {
br.close();
catch(IOException ioe) { ...}
}

Re:Java - the perverse language

[pete-classic]

You mean Java doesn't magically close the stream when it goes out of scope?

(I say this not as a lazy programmer, but as someone who feels that Java breeds lazy programmers.)

-Peter

# downloads mean little

[jeffy124]

The number of downloads means little here. I bet that many of those downloads were people looking to evaluate the various J2EE app servers. I know from experience that the vendors like BEA and Oracle do make free trial licenses available for people in the process of picking their app server.

The interview points to a survey among the sales of various J2EE app servers, and correctly points out that JBoss has no sales to contribute, making the survey be the data among commercial app servers.

clearly the best way to find the JBoss market share is to survey those who use J2EE, not just the companies that make them.

Re:# downloads mean little

[Delirium Tremens]

The number of downloads means little here.

Dude, you are out of your element here. 70,000 downloads per month means little? Or you mean, this is happening because there is one single dialup user who keeps retrying the download because his 56k modem has a baud fart? Yeah... right.

Re:# downloads mean little

[richardoz]

Bea (Weblogic 6.1), IBM(Websphere 4.0) and Oracle (9iAs) are all available for trial downloads.
BEA
Oracle
IBM
I recommend Jboss to my customers..

Re:# downloads mean little

[Delirium Tremens]

You should actually recommend the one from Borland. It has the best CMP, IIOP and security support I have seen so far.

Re:# downloads mean little

[Anonymous Coward]

That's new stuff from BEA. As late as last year, they made you kiss the salesman's ring before you could get an eval copy.

Re:The Turd Report 11/29/2001

[Anonymous Coward]

i bet if for a few days you ate nothing but broccoli, asparagus, and peas doused with as much green food coloring as your body could take without being poisoned, you would have smelly, watery, green diarreha for about a month.

Subset of JBoss capabilities...

[sterno]

Actually JBOSS includes Tomcat within it's framework for serving up Java Servlets, JSP's, etc. Tomcat is purely a server for Web Applications where as JBOSS does EJB's, and a whole slew of other things.

Real Use

[peterdaly]

from their web page:
JBOSS SUCCESS STORY OF THE MONTH! NEW

"Just through you'd like to know that the United States Department of Labor's Office of the Chief Financial Officer uses JBoss to process about $3.0M worth of financial transactions yearly in one application alone. There are several other legacy applications scheduled for migration. By using JBoss, we've saved the taxpayers about $100,000 in BEA Weblogic licensing fee and about $10,000 in annual support fees".
Michael R. Maraya, OCFO/OFD/DFAD

Christ's sake...

[Tom7]

Can anybody actually read more than a paragraph of that shit without their eyes glazing over?

I can read scholarly papers about programming languages in all their greek letter glory, but this is just too much. It's all enterprise-this and acronym-that and just terrible to understand. Maybe I should have gone to business school? ;)

Re:Christ's sake...

[mrcparker]

I agree. It seems like being a Java developer requires not only knowledge of the language but knowing about 10,000 extra acronymns also.

Re:Christ's sake...

[easter1916]

Yes, it does. There's a huge base of libraries and APIs out there, each with an acronym to make it less intelligible. Pity, because the underlying software is usually excellent.

Re:Christ's sake...

[jilles]

That applies to most languages. Just understanding the syntax is about an hour or so of work for most programmers (shorter if they are experienced, longer if they haven't seen a real language yet).

However, whether you are dealing with the STL, MFC or the Java API, it takes time to get productive with that. The Java API is simply huge and covers an enormous amount of functionality that you don't have to invent yourself. Luckily, there is JavaDoc that helps to browse through APIs. In addition, most APIs are well designed and have been through extensive review processess (unlike the crap Microsoft pours out over you) so if you know your design patterns, you won't have trouble understanding how everything fits together.

If other languages appear to be simpler, maybe that is because they don't include so much functionality and require you to do more work reinventing the wheel.

Re:Christ's sake...

[Anonymous Coward]

Doesn't Stanford's Business school have an "Acronmys for PHBs" MBA program with sub specialties in SUN/JAVA acronyms and/or MSFT acronyms?

Re:Overkill for anything that's not mission-critic

[dildofire]

but for simple web based apps, which are (or at least should be) relatively simple applications, you don't need a "real" programming language. perl gives you all the functionality you need.

i work for a company that chose to do it's web based ui in a "real" language (java). our developers got so caught in using all the whiz bang OOP features of java, that they didn't solve the simple problem that was put in front of them. java is great for certain apps, but for simple web/db apps it's overkill.

Re:Welcome to the Hot Squishy Submarine

[Anonymous Coward]

you cock sucking motherfucking whore, you shall die as pour sulfuric acid down your deposit slot

Re:The Turd Report 11/29/2001

[Anonymous Coward]

Make jello shooters out of lime koolaid, it'll make your turds bright green.

You missed a layer

[sab39]

You missed the most important layer, and I'm betting that that's because that layer actually isn't open source.

Are you running this on Kaffe? gcj? ORP? Kissme?

Didn't think so. (If you actually are, I'm dead impressed - please let me know how you managed it)

You're using Sun's J2SDK. Which isn't open source.

I'll be very happy when it really is possible to put together an open source J2EE stack. But that day isn't today, because the VM/classlib layer has no open source alternative that's up to running these enterprise-level apps.

Re:You missed a layer

[adamy]

Touche'

I really should try it with one of the open source compiler and VM combinations, but I haven't yet.

I was going to say that believe the Sun JDK for Linux is based on blackdown's work, and the blackdown stuff is Open Source, although not GPL. However, I just checked the balckdown site and the code does not seem to be available.

Re:You missed a layer

[Anonymous Coward]

It's not even Open Source as defined by Open Source Initiative

Socialist Marginal Characters

[JamesOfTheDesert]

I'm all for people feeling good about their work, but Mr. Fleury is a bit bombastic abut his product. Makes it kind of hard to really believe what he's saying (and oddly similar to MSFT sales pitches).

If he really is planning on something called JBoss.net, I think he'll find that 1) people expect it to run MSFT .net code, and 2) MSFT will expres an unpleasant interest.

And just what are "socialist marginal characters"? Are they for or against "WebOS" monopolies?

And why is SUN in all caps? Is that to remind people that it was started from publicly-funded research (Stanford University Network )?

Re:Socialist Marginal Characters

[easter1916]

Hey, he's French and so is prone to hyperbole. It's still a great product.

JBoss is a great way to add EJB container support

[MarkWatson]

I suspect that most people writing web services in Java can get by very nicely just using Apache Tomcat to provice servlet and JSP support.

That said, the EJB component model is cool, and getting the Tomcat/JBoss bundle is a great way to get started with servlets/JSPs/EJBs.

To add 'icing on the cake', add the Apache SOAP servlet (or the newer Apache Axis servlet) and also get SOAP support for interop with other systems.

- Mark

PS. I live in the mountains in Northern Arizona, and it is snowing. Cool :-)

We're switching to JBoss

[under_score]

This isn't a big deal, but I've been very wary of J2EE app servers. I've been working with server-side Java for about 4 years now, and previously with EOF (Nextstep/Webobjects). EJB is incredibly broken compared to EOF/WebObjects. But my online educational system Oomind is running on Tomcat right now, and needs a more complete platform. So we are moving. And after some careful analysis and real world experience, we're choosing JBoss. Frankly, it being open-source is a very significant factor. Kudos to JBoss and its developers!

Market shares are not of use, but of sales

[icoloma]

What you are talking about is that Weblogic and Websphere together are 50% of app servers _sales_. Not installations. That's why JBoss didn't even appear in these statistics.

JBoss is free, and has the same problem of linux - you just can't know how many installations are out there, only guess based on the number of downloads. If you look this, Jboss has by _far_ hundreds of times more installations than Weblogic or Websphere - in a month.

And, about the certification areas: they have complete certification courses, that are said to be of *very* high level. That being said by some java certified architects.

Changing to JBoss from other App Servers

[searlea]

2 years ago I had to do an evaluation of WebLogic, Oracle's JServer, and NAS/IPlanet (i.e. the last NAS version and the first IPlanet version.) What was annoying at that time was that there wasn't really a standard deployment method (I'm not even sure if there was an 'ear' spec.)

What I'm wondering is; has it all settled down now? How hard would it be to take EJB's developed and deployed for WebLogic or WebSphere, and to move them into a JBoss environment? Obviously if you use specific features of the app servers that's an open ended question - let's assume we're talking about sticking to the J2EE spec though, only using standard J2EE API's?

I'm hoping the answer's going to be very easy. I'm currently looking for work, and people keep asking if I've got WebSphere experience. I keep replying that it's just another Application Server... I can't imagine there being any surprises. Am I wrong? There don't seem to be many web design roles asking for IIS/Apache differentiated experience, so why should an EJB job differentiate between Application Servers?

Re:Changing to JBoss from other App Servers

[maeglin]

There are still differences between servers, but it's getting better. If you start working on WebSphere the are a few differences in the way the EJBs are written as well as how they are deployed vs. something like WebLogic but if you're using Visual Age for Java (which makes EJB creation a breeze) that's not your biggest problem. WAS 3.0-3.5 is a colossal PITA to work with without blowing it up.. if a job offer is involves any app server support, run away. WAS 4.0 has improved the compatibility and it's got a better EJB builder so you can whip up a new one quickly even if the old one doesn't work but the maintainability, although a lot better, still leaves a bit to be desired.

Re:Changing to JBoss from other App Servers

[FatherOfONe]

To answer your question about how easy it is to develop for different application servers. It is easy. The code ports great. Now that is not to say that deploying your EJB will be easy. Each app server has it's own weird way of doing that.

To answer your question about everyone wanting Websphere experience... A lot of shops use IBM mainframes and IBM is pushing very hard with JAVA and Websphere. Soooo upper management says that they need WebSphere guys. Understand that there are two types of people that they may be looking for
1. Developers - If you have done EJB development, you should have no problem. Although using Visual Age is a bit different.
2. LAN Guys - These are the guys who will maintain/tune the box. If this is what they are looking for then you should know Websphere and the O.S. it runs on, and it makes sense that they want someone who knows that particualar product.

Re:Changing to JBoss from other App Servers

[Kerg]

To get some idea where things are going currently, you can check this article from JavaWorld:
The Art of EJB Deployment

You don't know?!!

[Breace]

In case you have been living under a rock, JBoss is an Open Source Java Application Server (J2EE)

That's right. Don't you just love it when people think the entire universe revolves around what they are doing.

But both you and I will be flaimed for being so ignorant. Even though it's pretty busy under this rock.

Read the interview; what I can't figure out is...

[HermanH]

...why the banner ad at the bottom of their page says, "TO WASTE MONEY...SWITCH TO JBOSS...IT'S UN-AMERICAN." ;-)

Re:Overkill for anything that's not mission-critic

[easter1916]

It sounds like your project management team wasn't very good, then.

Horizontal scaling?

[MSBob]

One of the huge drawbacks of the current JBoss package is the lack of the ability to distribute your beans across multiple phisical hosts. That makes JBoss applicable to only small production environments. I wonder if this is something that 3.0 is planning to address... If not then JBoss will not be able to compete in the same space as WebSphere and WebLogic.

Re:Horizontal scaling?

[xnuandax]

JBoss 3.0 supports entity and session clustering...

Re:The Turd Report 11/29/2001

[avandesande]

Did you just get in a fight? Usually a black stool indicates that you have internal bleeding.

Environemnts

[SuperKendall]

Sadly, WebSphere is the one area where I have no experience. But I have used iPlanet (up to 6.1) and WebLogic 6 for a good deal of time, migrating an app from iPlanet to WebLogic...

Basically, the code ports really well and with few modifications. But where the differences come in are development environments and deployments, which would probably be different from one company to another anyway even using the same app server.

Another area where you can find differences is in frameworks used - what people actually do on top of the EJB standard can vary quite a bit while still being compliant. Then you have add on frameworks for front end, things like Struts, which aren't really part of the spec but are widely used.

What I would do in an interview situation is describe how the websphere deployemnt you worked with was set up - what you used to develop, how you deployed, etc. Then ask them questions about thier deployment and see how it compares.

I love this quote...

[TFloore]

Odds are the product with the greatest market penetration will always have the greatest stability.

(From the last paragraph of question 3.)

Microsoft would be the exception that proves the rule?

I tried to resist saying that, really I did...

Question about clustering

[RelliK]

I hope somebody can answer this. I am wondering how the application-level clustering (i.e. what JBoss, Weblogic, et al have) compares to the IP-level clustering (i.e. something like the Linux virtual server, or the embedded hardware equivalent from Cisco et al). There would be less overhead at the IP level, so, it seems the load balancing would be more efficient. Also, by doing load balancing at this level you can cluster just about any application -- be it a web/ftp/file or, as in this case, application server. What's the advantage of having clustering built in to the app server? How about Apache mod_backhand?

Re:Question about clustering

[Anonymous Coward]

Load balancing done at the hardware level is the most efficient. However, when it comes to applications, you usually have some state that needs to be synchronized between the different component hosts (unless you're able to design your application completely stateless, or are happy to synchronize the state at the DB level).

Anyway.. the state synch needs to be done at the app server for the load balancing to work. Failover as well.

Re:Question about clustering

[Marsh Jedi]

Session-state peristence.

2 Scenarios:
1. w/o clustered persistence:
You have two appservers. One dies. You fail over, the app server has no idea who you are. Log in again. What state was your transaction in again? Did you buy that widget or not?

2. w/ clustered persistence:
You have two app servers, replicating session between the two. App1 fails. App2 knows about your session, picks up where the other left off.

Re:Question about clustering

[RelliK]

OK, so is app-server load balancing a substitute for IP-level load balancing or is it supposed to be used in conjunction? (i.e. something like balancing the web servers with a IP-level load balancer and replicating the application state with app server load balancing?)

Re:Question about clustering

[JamieF]

There's a more specific benefit to clustered persistence via application-level session replication: less load on the database. Allow me to elaborate.

For a stateless app, such as a public web site with static content, IP based load balancing works fine. For a stateful app, there are actually 3 options, not 2 as mentioned above.

1) Neither save nor replicate state. App dies, user fails over to another server, back to login. Lame.

2) Save state in the database after every page view. App dies, user fails over to another server (doesn't matter which one), other server notices it doesn't have the user's state in memory and loads it. Acceptable, but complex, and beats up the database quite a bit. If I am not mistaken, this is the only option for most non-Java app server situations (CF, ASP, PHP).

3) Replicate state directly from app server to app server. After every request, the app server copies the state info for each user to a failover server. Server dies, load balancer redirects the user to another server, that server has the state info in memory, ta da, no loss of state, and no load on the DB. The new app server starts publishing the user's state info in case it also dies.

For a concrete example, imagine an e-commerce app with a shopping cart or complex product search as your main state. In #1, you lose your shopping cart and get logged out; log back in and your cart is empty. Suckage. In #2, it works but the developer has to save the shopping cart or search criteria to the database every time the app services a request, because any request could be the first one that goes to the failover server. OK for you, super suckage for the developer, and the DBA, because every time you ask the server to do something, it does that and then also saves your shopping cart, so the DB is constantly doing updates to a single table no matter what else it's doing. (In case you don't know about databases, writes are far more costly than reads for an RDBMS, so it's much more work to save a shopping cart again and again than it is to fetch product info, show all products whose vendor name is 'Sony', etc.) In #3, it's pretty much transparent to the developer, and the DB is just doing the work you would expect would be needed based on user activity. You just need a slick app server to make it work.

Clearly, keeping state info small is necessary to prevent session replication from eating up all your bandwidth. Also, every app server I've seen which does this picks a single backup app server to send session info to (rather than copying each session to several other servers each time). This means that the load balancer must choose the right server to failover to; this is done by encoding the IP address of the primary and backup app server in the session cookie, so the load balancer can use that info as it recieves a request to decide which server will get the hit. When the backup server gets hit, it changes the cookie to say it's the new primary, and to say which new IP is the one for the new failover server. If I understand correctly, in an imaginary server farm of 8 app servers, each user might have a backup session on any of the other 7 servers - the app servers choose this dynamically.

Anyway, the big idea here is that you tell all the app servers what all the IP addresses are for the other app servers, and keep your session object small, and It Just Works. No need to copy junk to the DB all the time; no need to check if you have the info cached. But your load balancer has to support your app server so it can do the "session affinity" thing - returning a user to the same app server again and again unless the server dies.

Re:Question about clustering

[Marsh Jedi]

Clustering != load-balancing.
Two completely different ideas.

Load balancing is simply splitting requests between multiple servers. Because HHTP is a stateless protocol--it remembers nothing between GET or POST requests, every time you go to a site, the webserver does not know you from Adam. This really sucks when you are trying to get fifty pieces of information from someone, put it in a database, and then _inform_ that person authoritatively that it went into the database.

The match made in hell between the stateless HTTP protocol and _very_ stateful, transactional datastores gave birth to an entire industry--middleware: arbitrary server-side datastores that could keep track of a user's session info, keyed on UIDs in cookies, or written in the URL.

When they tried to round-robin between their new, spiffy, stateful system, they realized that because only one server instance had the session info, the chances of hitting the same server were pretty slim.

So they began replicating the session-state between the servers, and called it clustering. To this day, it sucks and doesn't work--under load, the app server might not have time to replicate to the others in your cluster, because software is _slow_. What it _is_ good for, tho, is failover. Odds are the primary managed to at _least_ send your session-state off to the secondary before dying.

Finally, vendors like Cisco they started putting sticky-bit features in their hardware load balancers (read: LocalDirector) that make sure requests from a given IP "stick" to the first server they hit--replication is generally used for failover.

Re:Question about clustering

[jrumney]

For load balancing only, IP level clustering works great. But to allow failover in an application that needs to maintain some state between requests, the app servers also need to be clustered. It is not a competing technology, generally you would use both IP level and app-server level clustering together.

Re:Question about clustering

[Anonymous Coward]

I'll take the second option. In the third choice your app server has to copy junk to one or many other servers in the cluster to keep the sessions fresh just in case it dies. In the second it copies junk to a DB. Saying that writes are too slow to use on a DB ... why have a database then? Strictly for reads? If your app server can't trasparently do app clustered in-memory sessions AND db/db-clustered persistance... what good is it?

Oh yeah, the second option doesn't require sticky load balancing.

Re:Overkill for anything that's not mission-critic

[chromatic]

You're probably thinking of PERL. As nearly as I can tell, PERL is a dumbed down version of Perl (see the difference?) used for writing buggy, insecure message boards and spam gateways posing as mailer forms. It has as much in common with Perl as JavaScript does Java, circa 1995.

Re:Read the interview; what I can't figure out is.

[Anonymous Coward]

I think that it is a reference to a statement by MS saying Opensource was un american or something like that

Re:PHP /MySql vs. multi-tier apps

[igbrown]

Well, you would probably not want to replace a PHP/Apache/MySQL deployment with a full-blown J2EE application server. The model you describe is database connectivity and business logic being handled by the web tier. This would map to a Apache/Tomcat using JDBC (Java database connectivity) embedded in servlets or JSPs. While this technically IS an application server a more practical design is using Apache\Tomcat to handle web requests from a browser, and then in turn access an application server with the appropriate method call or database query. This removes logic and database calls from what is essentially a publishing platform (PHP|Tomcat+Apache) and isolates these functions in a single location. The main benefit is that you can add new types of clients (a standalone application of some sort for instance) without have to recode logic or database querys that would optherwise be embedded in JSPs, servlets, or PHP.

The gist of it is that you get a cleaner separation between logic, data, and presentation.

Re:Read the interview; what I can't figure out is.

[Anonymous Coward]

"switch to JBoss...it's un-american...to waste money."

Same discussion on serverside.com

[jeet_means_win]

There are some better post at http://theserverside.com ... a few interesting comments from Rickard Oberg http://www.theserverside.com/home/thread.jsp?threa d_id=10469

Re:Same discussion on serverside.com

[Anonymous Coward]

Does anyone have a copy of Rickard's anti-US diatribe he gave JavaLobby.org on September 12th? It appears JavaLobby.org has removed it. It certainly has changed my mind about using JBoss (or should I say TalibanBoss).

Re:Same discussion on serverside.com

[Anonymous Coward]

For the record, my diatribe was about anti-western media, not anti-US.

And if that kind of completely unrelated stuff makes you change your mind about using JBoss, then you're probably not the kind of user who would appreciated it anyway ;-)

/Rickard

Re:Same discussion on serverside.com

[Anonymous Coward]

For the record, my diatribe was about anti-western media, not anti-US.

Yes, on this point we agree. It seemed you were in favor of anti-western media as I recall. You also had some very interesting theories on September 11th that Americans would find very interesting. Please set the record straight and post your JavaLobby comments on the issue here.

Re:JBoss is a great way to add EJB container suppo

[clarkgoble]

Cool. That's a beautiful area of Arizona - about time your ski hills could open. Here in Utah it's been snowing for a week. (About time for us as well, given the past few years)

Re:java

[Anonymous Coward]

cs is more fun when you get your hands dirty.

I don't get it. What makes counterstrike so fun with dirt on your hands?

Seems like you'll just mess up your keyboard.

IIOP

[Anonymous Coward]

I happen to believe that the biggest obstacle is the lack of RMI/IIOP wire implementation in JBoss and I do believe some important person in the JBoss development team hates the idea. Without the RMI/IIOP, every wire payload for security and transaction is destined to be proprietary. Weblogic got t3 RMI, Orion (aka Oracle ;-)) got ORMI, etc. They are not interoperable with anybody else.

I can always get a powerful machine to run a single instance without the distributed replication headache in clusters. However, I just can't do transactions/security/failover using interceptor in a portable way around those application servers. In this regard, I do believe that OpenORB and OpenCCM are more honorable projects.

I don't like to be locked into any "vendor", open source or not. Unless JBoss is sincere about the ultimate interoperability thing, they are unlikely to be taken seriously.

On the other hand, their commitment to a well organized JMX architecture is very commendable.

Re:IIOP

[Kerg]

RMI/IIOP is in the works... not sure if it will make it into 3.0 but shouldn't be too far off.

What are the free IDE choices

[Anonymous Coward]

I use WebLogic and JBuilder on NT and Solaris at work--my company likes to spend money and then complain they're broke. If I were developing at home, I would obviously pick free products like Apache, Tomcat, JBoss, and MySQL.

But what do people use for a free IDE? Is JBuilder really free for home use? Given my experiences with the JBuilder salesman--he wanted full price for 5.0 when we had bought 4.1 a month before--I find this hard to credit.

Re:What are the free IDE choices

[zeppelin71]

haha - well they just announced JBuilder 6 :)

Try www.netbeans.org for an open source java IDE... its sponsored by Sun, and forms the base code for Sun's own commercial offering ... Forte.

Many banks use multiple server platforms

[beroul]

I work at one of the largest vendors of software for banks, and diversity of server platforms is a fact of life for our clients. Many banks are quite happy to continue using applications that were written 20 years ago, in RPG, for the IBM AS/400 platform. And it makes sense; those applications work quite well, and they represent millions of lines of code, which can't be thrown away just because RPG isn't a modern language.

Many of those same banks also have applications running on all sorts of Unix systems. And of course, Windows is widely used for front-office applications. Oh, and they all use different RDBMSs, too.

If you think it's easy to write, say, C++ applications that are portable across all those environments, I suspect that you've never really tried.

We've written our in-house middleware in Java; we develop on Windows desktop machines (alas!), and the code works out of the box on AS/400, Solaris and Linux, as well as with every RDBMS we've come across. It's an easy sell to our customers, many of whom have strict requirements about which operating systems they're willing to use for which projects. And we haven't spent any time porting our software.

Overly optomistic?

[gss]

While I am a big fan of JBoss (ever since EJBoss), it does have a long way to go to gain market share in the industry. People who make the decisions would rather put their neck on the line for big companies like BEA or IBM, be it right or wrong. Does this guy really think that BEA and IBM aren't working on their next versions as well? It's a tough market, I just hope JBoss doesn't get too confident.

Cool Project, but JBOSS has some issues.

[Anonymous Coward]

JBoss is a cool project, but Fleury has some ego issues to deal with. In general on the mailing lists I found his tone towards new users to be aggressively insulting at the least.

Additionally right now they are discussing pulling down the documentation to force people to buy their books or pay for access to the documentation. This coupled with his refering to a Debian developer as a "Childish dreamer" and being "Shit out of luck kid." (just because they wanted the non-free, non opensource software bundled with the source put in a seperate package) makes me think a bit ill towards the whole project.

The good part is the code is LGPL so unless he calms down the code could easily be forked. (Which might make sense seeing JBOSS Inc's less then great idea on making money off of documentation. As soon as they pull the docs someone will quickly post up a new free version that will probably be better than what currently in the SAMS book.)

Re:Cool Project, but JBOSS has some issues.

[Anonymous Coward]

If I remember correctly, the reason he slated you was because you started slabbering at some of the developers on the list, like a total dick

Keep *your* ego issues in mind.

Re:Cool Project, but JBOSS has some issues.

[Anonymous Coward]

Actually I'm not the person in question, just another debian user wanting a functioning package.

Want a performance tip?

[tgd]

Try JBoss+Jetty in place of JBoss+Tomcat running through Apache. Unless your application is a very simple one, you'll get a 2-4x increase in performance under load.

Tomcat is nice, in theory, but its dog slow.

Re:Server-side Diversity? - hot just in one org...

[Anonymous Coward]

Yeah, people might not WANT to mix platforms, but it's often a necessity. But more to the point, you often need to interact with other organizations who may not run the same platforms as you, and to be able to tell them that they can still run the same package as you are for consistency sake is a good thing.

Awesome.

[Futurepower(tm)]

Awesome.

Wow, if Microsoft had said this....

[GCP]

Wow, if anyone tells me that open source is better because it doesn't have the hype of commercial products, I'm going to hand him a copy of that article. That should clear up that little delusion REAL fast.

"features that strike fear in the hearts of our competitors...have made us the darling of serious development shops. ...the rest of the industry is still scratching their heads to figure out what this really means. Some follow suit, (BEA/IBM), others just give up. But we keep on trail blazing tech-wise...we are going orbital..."

If Microsoft talked like that, you guys would roast 'em.

"Odds are the product with the greatest market penetration will always have the greatest stability...."

Hoo, Mama!

Re:Wow, if Microsoft had said this....

[Anonymous Coward]

Oh come on. Give the guy a break. JBoss clearly occupies an unusual position between the corporate and Open Source communities. Neither fish nor fowl. The corporate world doesn't quite know what to do with them because they're Open Source (and not VC-funded or sponsored by some big corporation) and the Open Source community is equally wary on their end because JBoss is an Open Source project delivering Enterprise software. (These are the people bitching about paying $10 for the advanced documentation, since it doesn't occur to them that you might have to provide some financial motivation to get something like that done).

Kudos to Marc and JBoss. You have to admire them for getting where they are today. One thing he clearly does is give credit to and promote his developers, something not common in the technology companies, God forbid you work in the IT dept. of a non-technology company.

Aiya!

[samskivert]

That was about the most annoying thing I've read since the last time I was foolish enough to peruse a Microsoft press release.

Perhaps someone should remind Mr. Fleury that people in the "open source" world are interested in the technical details and not interested in hearing about popularity, success or prospects for ruling the world. We can make the latter assessments on our own, thanks very much.

Re:Aiya!

[Anonymous Coward]

I believe the point is to garner the interest of business world, not especially the Open Source world. The OSS comes along, have come along because the source is there, its technically advanced and interesting. That alone has made it interesting for the OSS world.

Now you need businesses to realize they can use OSS too. Convincing the IT managers is not easy. Having Microsoft-like press releases might just work, who knows. When the project has come as far as JBoss has, it already has its own momentum in the OSS world.

Opensymphony Plug

[farkas]

Another open source java tools project is Opensymphony.

• Oscache is a must for any JSP / servlet development, and allows the caching of sections of pages.
  
• Sitemesh allows you to decorate HTML similar to the decorator pattern described by the Gang of Four.
  
• Transform tags allow you to do code highlighting and xml transformations in your pages.
  
• WebCompass allows you to create meta-models of your site (allowing for dynamic navigation).
  

We use all of these features on our site.

Any ``prominent'' sites running Jakarta/JBoss?

[cygnusx]

Some questions from a guy who has played around with Jakarta a lot, but never quite used it at work:

Are there people who are using Jakarta/Tomcat for high-volume sites? I use Tomcat for development all the time (primarily because I can use it on some really low end hardware). But in production use, it's always been a different story.

I mean, I can point to /. when people say MySQL cannot handle loads -- would any of you know a similar highly-trafficked, prominent site that runs Jakarta software?

For that matter, how 'bout the same question applied to JBoss? or are they used quietly in departmental intranets, anonymous because they didn't require a big-ticket purchase order to acquire?

I want to know if you consider Jakarta/JBoss ready from 'prime-time' in the same way (say) Websphere is.

Re:Any ``prominent'' sites running Jakarta/JBoss?

[Enygma42]

We deployed a site on Tomcat that gets about 60,000 hits a day which I would say is a considerable amount. The thing is we put squid in front of Tomcat so it doesn't have to do very much. I've done tests between Tomcat and Resin (http://www.caucho.com $500) and Resin turns out to be consistently 5-10X faster. We'll be pushing clients to use Resin from now on.

Re:Any ``prominent'' sites running Jakarta/JBoss?

[Anonymous Coward]

Was that tomcat 4.x or 3.x that Resin beat? I understand 4.x is supposed to be a lot faster, but I haven't seen the benchmarks.

I've screamed at IBM (and BEA)

[Guppie]

In Norway, IBM is by far the most arrogant and useless company you can buy software from. My company have used Websphere and DB2, and the help from IBM (after hours of phone-screaming) was worth next to nothing.

We used Websphere 2.x and 3.0, IIRC, and they were CRAP, i know DB2 is a decent database, but we were not able to configure it to run any faster than mySQL, thanks to lousy support from IBM.

In the end, we dropped Websphere for Weblogic, no great support from BEA either, but at least the appserver ran 24x7 w/o great problems...

Re:Cool Project

[Anonymous Coward]

will probably be better than what currently in the SAMS book

Heh. And I wish that person all the luck in the world. You don't have no fucking clue how much work it is to put together a high quality documentation for a product so big as JBoss.

Because you're just a dick sucking OSS whore.

72,000 whole servers???

[Anonymous Coward]

AFAIK, jboss is not a packaged product but more of a composite server (the container itself, JMX, messaging, queuing...) I guess that you'd need 10s, possibly 100s of downloads in order to have a production-ready framework.

Also, this is open source. Each sub-project is constantly improved and debugged; that's the beauty of OSS. A single "customer" probably downloads again the whole stuff every month. If you deduct contributors to the project and people who d/l just for testing purposes, i guess that the number of new "customers" is several orders of magnitude below 72,000 per month.

You might still get honest figures; but please, don't believe jboss takes $10,000*72,000 or 720 millions of sales from BEA each month. Just because it's OSS doesn't mean we should fall for lame marketing a la M$.

Re:72,000 whole servers???

[Kerg]

AFAIK, jboss is not a packaged product but more of a composite server (the container itself, JMX, messaging, queuing...) I guess that you'd need 10s, possibly 100s of downloads in order to have a production-ready framework.

Uhh, no you don't. You'd find out this yourself if you ever actually downloaded the product. It's one download, two if you want to have both servlet engines prepackaged.

Between Open Source and Corporate world

[Anonymous Coward]

We read a lot about the corporate adoption of Linux. JBoss is critical to this, because it's "the application stupid."

Do you realize that any application developed on this framework can be deployed on Windows, Solaris, AIX, and...Linux. This really gives an entree to Linux in the higher spheres of Enterprise computing--the holy grail of commercial computing. It's not with cute GUIs that you crack the Enteerprise market. It's with rock solid infrastructure software like JBoss.

From Marc's interview, I also understand that they are trying to come up with a commercial model for Open Source, one where developers come first. Contrast with Linux where third party packagers fight for meager earnings and Apache which killed its own market and is today co-opted by IBM/Sun.

Rickard's September 11th diatribe

[Anonymous Coward]

For the record, my diatribe was about anti-western media, not anti-US.

And if that kind of completely unrelated stuff makes you change your mind about using JBoss, then you're probably not the kind of user who would appreciated it anyway ;-)

Rickard


Rickard, can you please post your "September 11th" JavaLobby comments here? I am sure Slashdot readers would love to hear them. Please don't exclude your remarks about Isreal and how you feel the U.S. somehow deserved this tragedy because of their foreign policy.

Re:Cool Project

[Anonymous Coward]

Actually having documented projects to military specification I probably know a *LOT* more about it than you. Take a clue, Fleury owns the Trademark JBoss and is trying to push a non-open, non-free, opensource project.